Gov. JB Pritzker signed SB 315, the Artificial Intelligence Safety Measures Act, on July 6, making Illinois the third state to regulate frontier AI developers and the first to require annual independent third-party safety audits. Combined with California and New York, the three-state bloc now sets rules for roughly 40% of the U.S. AI market. That’s the de facto federal policy while Congress hasn’t moved.

The law targets “large frontier developers”, those with at least $500 million in annual revenue, and takes effect January 1, 2027, with enforcement handled by Attorney General Kwame Raoul’s office. Companies must publish safety frameworks addressing “catastrophic risk,” defined to include death or serious injury to more than 50 people or $1 million in property damage. Incidents must be reported within 72 hours, or 24 hours where there’s imminent risk of death or serious injury. Maximum fines run to $3 million per infraction, which Raoul called “a beginning step.”

The audit provision is where Illinois breaks from the pack. New York’s regime required only a single independent audit when a developer first qualifies. Illinois wants one every year.

Politically, the vote wasn’t close: 110-0 in the House, with only five Republican no votes in the Senate. OpenAI and Anthropic backed the bill, and Anthropic sent representatives to the signing. TechNet objected in May committee testimony, with representative Ninia Linero warning that audit requirements absent national standards would produce “highly subjective determinations.” House sponsor Rep. Daniel Didech, D-Buffalo Grove, cited Anthropic’s own Mythos model, which the company withheld as too potent a cyberweapon to release, as evidence that harms “are not theoretical.”

The larger structural read: the frontier labs that could most easily fight state regulation are instead endorsing it, because a patchwork of three friendly states is cheaper than a hostile federal bill they can’t yet see. Colorado and Texas are the next dominoes worth watching.

Sources

Sources